Free audits are sales calls in a lab coat: how to evaluate a claims-recovery vendor

At some point every practice owner gets the email: we will audit your denials for free, no obligation, you will be shocked what we find. Here is the thing about that offer. The audit is not the product. The audit is the sales call. A free audit exists to produce a number large enough to close you, and the person running it is compensated on whether you sign, not on whether the number is right. That does not make every free audit dishonest. It means the burden is on the vendor to prove theirs is different, and this article is the checklist for making them do it.

Why the free number always comes back big

Start with the raw material. The industry statistics genuinely are dramatic: per MGMA, 50% to 65% of denied claims are never reworked or resubmitted, and HFMA-aligned analyses find that practices not actively managing denials often have 5% to 10% of gross revenue tied up in denied or underpaid claims. A vendor does not need to fabricate anything to produce a headline figure; they just need to quote your gross denied dollars and let you assume it is all collectible. It is not. Some of it is past timely filing. Some of it was denied correctly. Some of it costs more to work than it will return, given rework costs that MGMA, HFMA, and Change Healthcare estimates put between $25 per claim and $118 to $181 per appeal at the high end. An honest recoverable figure is the gross number after all three haircuts. A sales figure is the gross number before them. Ask which one you are looking at.

The gross-versus-net-recoverable test

This is the single fastest way to grade a vendor. Advisory Board research puts roughly two-thirds of denied claims in the recoverable category. That is an industry-wide figure, across all denial types and ages. If a vendor tells you that 95% of your specific aged denials are collectible, they are either sitting on the best book of denials in America or they are quoting you a number designed to survive exactly one meeting. A credible vendor will show their discounting: here is your gross denied AR, here is what dies on timely filing, here is what was correctly denied, here is what remains, and here is our expected collection rate on that remainder with a range, not a point estimate. If the walkthrough from gross to net-recoverable takes less than five minutes, it did not happen.

Questions that separate operators from salespeople

Ask who actually works the claims: employees, offshore subcontractors, or software with a human fallback, and who is accountable when an appeal deadline gets missed. Ask what happens to claims they decide not to work: do those get reported back to you with reasons, or do they silently vanish, leaving you to believe the worked claims were the whole book? Ask how the contingency fee interacts with claim size, because a straight percentage gives a vendor every reason to work your large clean claims and ignore the long tail. That may be fine, but you should know you are agreeing to it. Ask for the reporting cadence and insist the report include claims attempted and lost, not just claims won. Any vendor whose reporting only contains victories is grading their own exam.

Then ask the question most vendors dread: what would make you tell me not to hire you? A real operator has an answer, because real operators decline bad engagements: books that are mostly past filing deadlines, denial mixes dominated by correct denials, volumes too small to cover setup effort. A vendor with no disqualification criteria has told you that the free audit has one possible output.

Recovery and prevention are different products: make vendors say which they sell

One more source of confusion worth naming. KFF research indicates 86% to 90% of denials are preventable, and some vendors use that statistic to sell front-end prevention tooling dressed up as recovery. Prevention is valuable (arguably more valuable long-term), but it does nothing for the denied dollars already sitting in your AR, and it typically comes as a software subscription with a very different cost profile. Recovery works the existing book; prevention reduces the future one. A vendor blending the two in one pitch is usually obscuring the economics of at least one of them. Make them price and defend each separately.

The test was never the price tag; it is the deliverables

Full disclosure: our own Recovery Audit is free, so this entire article applies to us. Here is our honest reasoning. The audit is a $500 engagement (a credentialed analyst plus an AI pipeline over twelve months of remit data), and it genuinely costs us money every time we run one. We give it away anyway, because we would rather show a practice the benefit of working with us than describe it: a board-ready report on your own denials makes a better case than any pitch deck, and if the report convinces you, we have earned the engagement honestly. What makes it an audit rather than a sales call is not a fee; it is everything this article told you to demand. Every deliverable in writing. The gross-to-net-recoverable walkthrough with all three haircuts shown. Disqualification criteria that actually fire. And a written go/no-go that is allowed to say no, including "your AR is clean, you don't need us," a report you keep either way. Hold our free audit to every standard on this page. Any audit that will not survive that scrutiny, free or paid, is a sales call in a lab coat, and any audit that will is worth scheduling regardless of what it costs.